Representation
Can brain-body regulation under naturalistic contemplative conditions be represented as a low-dimensional latent dynamical system amenable to real-time inference?
Applied Research in
Self-regulating intelligence
Cogitos Labs is an early stage R&D laboratory based in India. We are building new AI architectures and systems grounded in a single foundational principle: resilient, responsible, and value aligned intelligence requires intrinsic self-regulation — not external supervision. Our work begins with the study of adaptive self-regulation in biological systems.
Contemplative Neurodynamics Program
Our first R&D initiative, with help of probabilistic generative models and dynamical systems theory, is studying biological adaptive regulation under contemplative practices through synchronized multi-modal physiological recordings. The program uses practices drawn from Yoga — meditation, controlled breathing, and postures (asanas) — as its experimental context, to voluntarily modulate regulatory states, and answer below questions.
Transition Markers
What physiological and neural markers signify transitions between regulatory states — and can these be detected with sufficient lead time to enable intervention?
Transfer to AI
What computational principles underlying biological adaptive regulation can be transferred to AI architectures to make them self-regulating?
Target Applications
Our research forms the foundation for developing a class of closed-loop adaptive regulation systems, which aim to enhance resilience and alignment in both people and artificial systems by enabling them to autonomously monitor, adjust, and sustain goal-directed behavior.
Clinical Platform
A closed-loop regulation support system for clinical settings, leveraging full range multi-modal sensing including EEG, HRV, IMU/EMG and respiration. Targets stress disorders, anxiety, and attention fragmentation.
Consumer Wearables
A broader-reach self-regulation support product operating on a lighter sensor footprint, designed for general population use. The same underlying R&D — made accessible beyond clinical settings to address emotional dysregulation and reduced resilience at scale.
Self-Regulating AI
A new class of AI architectures that maintain alignment with values over time through intrinsic self-regulation — representing those values, detecting deviation, and adjusting behaviour accordingly. The longer-horizon direction of the program.
Working Principles
The core intellectual commitments that govern our empirical investigation, computational modelling, and architectural development
Intrinsic regulation
Systems maintain and correct their own state through internal mechanisms — not externally imposed constraints. This is both our scientific framing and our architectural objective.
Naturalistic observation
Data is collected under conditions that preserve self-organised regulation. Experimental design is subordinate to maintaining natural regulatory behaviour.
Regulation as inference
Brain–body regulation is treated as a continuous dynamical process. Models perform inference over latent states, capturing structure and temporal evolution.
Co-evolving design
Physiological insights guide model design; model outputs expose gaps that shape experiments; architectural requirements feed back into both.
Reusable infrastructure
Pipelines, frameworks, and protocols are designed as reusable components. Each phase builds on prior infrastructure rather than starting afresh.
Responsible governance
Research with human participants adheres to informed consent and IEC oversight. Physiological data is anonymised and encrypted. We comply with India's DPDP Act 2023.
Founding Team
Raja Valiveti
Co-founder
Leads R&D direction, infrastructure design, computational modelling, AI architecture, and systems development.
IIT Kharagpur alumnus. 25+ years in technology, leading development and management of large-scale distributed systems and enterprise AI deployment.
Aruna Yerra
Co-founder
Leads experimental design, participant coordination, physiological data collection, and data governance.
PSG College of Technology alumna, Masters in Applied Electronics. Trained practitioner of Yoga.
Collaborate with us
We welcome conversations with researchers and organisations interested in the program's scientific and applied directions. Write to us about your interest and area of expertise, we will be in touch shortly.